Abstract: A large literature in political science and economics has emerged in recent years on the causes of corruption. This research has generally focused on either the economic factors that create opportunities and demands for corruption or the political factors that can reduce corruption including dense social networks of social capital, increasing legal mechanisms for monitoring and prosecuting corrupt officials, and political institutions that create electoral accountability by increasing competition and clarity of responsibility. The general consensus of the latter literature is that political competition allows voters to punish corrupt officials. We argue that corruption often exists as part of a larger clientelist framework where politicians tolerate the misuse of state funds by officials in order to gain access to the networks maintained by those funds. We thus build on recent work by Bueno de Mesquita et al (2000), Chhibber and Noorudin (2004), and Singer (2005) that argues that as the size of the necessary winning coalition decreases, political parties will emphasize clientelistic and personalist appeals instead of the provision of public goods and efficient economic outcomes to maintain their support. As a result, we argue that corruption should be highest in countries with either extremely high or low levels of electoral fragmentation. We test this proposition within the frameworks and datasets developed by Treisman (2000), Persson and Tabellini (2003), and Gerring and Thacker (2004). ..PAT.-Conference Proceeding [ABSTRACT FROM AUTHOR]